Tables to dataframes includes useless columns

[dilpath]

e.g.

>>> petab_problem.parameter_df
             lowerBound  upperBound nominalValue estimate priorDistribution priorParameters  index
parameterId                                                                                       
alpha               0.0        15.0          1.3     true                                        0
delta               0.0        15.0          1.8     true                                        1

In general we could omit empty columns and index.

This would mean petab_problem.parameter_df could be used in notebooks to nicely display the table, without custom formatting to remove the useless columns.

Also means that a user can round-trip their minimal PEtab tables from/to disk.

[dweindl]

I think there was a request elsewhere to always include all standard columns in the same order. There are good arguments for either. Ideally, we make it easy to switch between the two modes.